The Scorn of Humpty Dumpty (Part 1)
In a previous post, I summarized the objections I begin to raise here. This series will serve as a longer and more developed critique of the claim that there are only three forms of antirealism, and this post serves as an introduction.
One of the stranger claims in Huemer’s Ethical Intuitionism (2005) appears on pages 4-5, where he claims that there are only three forms of antirealism:
(1) Noncognitivism
(2) Subjectivism
(3) Error theory
Not only does Huemer claim that these are the only forms of antirealism, but that they are the only possible forms of antirealism:
“it is important to see why the above three positions are the only possible forms of anti-realism” (p. 5).
This is a very strong claim, and if Huemer offers a detailed explanation anywhere in the text why these are the only possibilities, I couldn’t find it (which isn’t to say it isn’t there). Page 5 asserts a series of dichotomies that seem kind of like an explanation for why these are the only possibilities, e.g., either moral claims do or do not purport to report facts, either they are or are not subjective, and so on.
Huemer seems to treat the only possibilities available to us as categorical, as though all moral claims must fall into one, and only one of these categories, i.e., either they are all propositional or none of them are, either they are subjective claims, or none of them are, and so on. But what if this isn’t true? I believe it isn’t, and that Huemer has overlooked alternatives to the three positions he describes.
Huemer seems to not consider the possibility that these may be false dichotomies it may be that the answer is “sometimes one, sometimes the other” (metaethical pluralism), “both at the same time” (also pluralism, but may result in incoherentism), and “neither” (indeterminacy). I don’t see any reason why these aren’t logically possible, unless, and only unless, we appeal to philosophical positions that would entail that these views are impossible that (a) go unstated in the text and (b) that I (and likely others) would almost certainly reject.
For instance, consider Don Loeb’s incoherentism. According to Loeb (2008), moral claims may commit speakers to conflicting commitments, e.g., there may be ineliminable elements of both cognitivism and noncognitivism, references to moral facts being both stance-dependent and stance-independent, and so on. If so, such positions would amount to attempts to say things like “there are square circles.” Such claims would be incoherent. This is somewhat different from error theory, both in terms of what problems it ascribes to folk thought/discourse and in terms of its semantic thesis (which is a type of pluralism). It is not error theory.
Yet Loeb’s position gestures towards the possibility of a broader and more fundamental disagreement about the nature of language. While incoherentism still attributes determinate (but conflicting) metaethical views to ordinary people, it points to the possibility that moral claims don’t necessarily fall into one or the other of the three categories Huemer describes, but could simultaneously fall into more than one. Though Loeb doesn’t emphasize this possibility, they could also fall into none of the above. It’s not clear why it is necessary that moral claims fit one of the three analyses Huemer proposes; why wouldn’t it be possible that they fit none of them?
In any case, we may ask an additional question: how can we know which of these analyses (if any) capture the meaning of ordinary moral claims? While Huemer and many other philosophers seem to employ the traditional non-empirical methods of contemporary analytic philosophy, Loeb claims that questions about the meaning of moral claims are empirical. I agree with Loeb.
If the meaning of ordinary moral language turns on empirical facts about e.g., what speakers intend to mean with those claims, findings could reveal significant pluralism in what people mean. If so, it could out that it’s just not the case that moral claims fit just one of Huemer’s three categories; it could be that moral claims fall into multiple categories, perhaps sometimes simultaneously.
Such pluralism may be sufficiently ubiquitous so as to indicate that no single analysis could adequately account for ordinary moral claims. If so, strictly speaking, (1)-(3) would have to be false, insofar as they intended to represent mutually exclusive analyses of moral language.
Notably, there has been empirical research on folk metaethics. We are approaching nearly two decades since such research began and earnest, and the results almost always reveal indications of variability. While I have serious methodological doubts about these findings, to the extent that we have evidence, it reliably points towards metaethical pluralism, not uniformity.
Of course, one could object that while pluralism would entail that no single account was correct about all moral discourse, it would still be the case that all moral discourse fell into one or more of the three categories outlined by Huemer. For comparison, if someone said “all fruits must be apples, bananas, or oranges,” this could mean that all fruits are apples, all fruits are bananas, or all fruits or oranges. It wouldn’t be that different to say that all fruits must fall into one of these categories, even if some fell into one category, and some fell into another, so long as it remained true that any given fruit was either an apple, banana, or orange (and not, e.g., a pineapple or a strawberry). If so, we could simply modify Huemer’s claim to say that all antirealist positions fall into one of these three categories, or some hybrid combination of these three categories.
Yet there is another possibility: metaethical indeterminacy (see Gill, 2009). Gill suggests that it could be that neither cognitivism nor noncognitivism does a better job of accounting for at least some moral claims, and the same could hold true for the other distinctions. If so, then in some instances there’d be no fact of the matter, or at least no discernible fact of the matter, about which analysis offers the best account of at least some moral utterances. It may be that moral claims turn out not to fit with any particular analysis at all. It may be, in other words, that moral claims don’t function to express the sorts of things philosophers have traditionally attributed to them.
To go back to the fruit example, this would be a bit like discovering a box of rocks, and being asked to categorize them as apples, bananas, or oranges. The answer may be “none of the above.” Alternatively, it may be that our observations are equally consistent with a given object being two or more different fruits, with no available considerations that would allow us to resolve the matter. While the latter might be better described as indeterminability, the former highlights a possibility that falls outside the scope of Huemer’s three categories.
Thus, while Huemer seems to believe these are the only three analyses an antirealist could endorse:
(1) Noncognitivism
“Murder is wrong” = “Don’t murder” (or some other nonpropositional claim)
(2) Subjectivism
“Murder is wrong” = “Murder is inconsistent with my/God’s/my culture’s moral standards”
(3) Error theory
“Murder is wrong” = “There is a stance-independent moral fact such that murder is wrong” and since there are no such facts, all such claims are false.
…There are at least three other possibilities
(4) Interpersonal & intrapersonal pluralism
“Murder is wrong” = some combination of (1), (2), and (3) This could vary between people (interpersonal variability) or with respect to different judgments (intrapersonal variability).
For instance, “murder is wrong” may express a proposition in one context, but express a nonpropositional attitude in another.
(5) Incoherentism
“Murder is wrong” = simultaneously some combination of (1), (2), and/or (3). Insofar as these are inconsistent such claims are incoherent.
For instance, a person who says “murder is wrong” may simultaneously be committed both to cognitivism and noncognitivism, in which case this claim is incoherent.
(6) Indeterminacy
“Murder is wrong” = Neither (1), (2), or (3). There is no determinate fact about whether it means (1), (2), or (3).
For instance, a person who says “murder is wrong” may have no position on whether they are referring to stance-dependent or stance-independent facts, in which case what they mean may be indeterminate with respect to this distinction.
All three of these possibilities exist in the literature by philosophers who specialize in metaethics and who work within the analytic tradition. Huemer claims only (1)-(3) are possible. I am not sure why. Is Huemer not aware of (4)-(6), or does Huemer not think they are legitimate possibilities? If so, why aren’t they legitimate possibilities? Why couldn’t there be ineliminable variation in the meaning of ordinary moral claims, as Loeb and Gill suggest?
I endorse (6). That is, I endorse folk metaethical indeterminacy (though I am receptive to the possibility that elements of 4 and 5 both appear as well). What I do not accept is what Gill refers to as the uniformity-determinacy (or UD) assumption: that ordinary moral claims have a shared, determinate meaning. The three possibilities Huemer outlines as the only possibilities all share the UD assumption, or, at best, do not explicitly reject it, which at the very least makes claims that these are the only three possibilities ambiguous: does mean one must endorse just one of these positions, or that one must believe all moral claims fall within one of these three categories, even if some fall within different categories? I don’t know. In my case it’s moot, since I think a substantial portion of folk metaethics is indeterminate, if I am correct, this at least represents one additional category that isn’t included in Huemer’s original three antirealist positions; one position remains: “none of the above.”
In the next post in this series, I delve into additional issues with Huemer’s claim that there are only three possible antirealist positions.
References
Gill, M. B. (2009). Indeterminacy and variability in meta-ethics. Philosophical Studies, 145(2), 215-234.
Huemer, M. (2005) Ethical intuitionism. New York, NY: Palgrave Macmillan.
Loeb, D. (2008). Moral incoherentism: How to pull a metaphysical rabbit out of a semantic hat. In W. Sinnott-Armstrong (Ed.), Moral psychology: The cognitive science of morality (Vol. 2, pp. 355-386). Cambridge, MA: MIT Press.